Dynamics of trust in collaborative complex system design

Abstract

This paper examines how trust is constructed, eroded and rebuilt in collaborative complex system design in safety-critical settings. The paper uses a qualitative process study of a 13-partner European project developing mid-air collision avoidance systems (MIDCAS) for remotely piloted aircraft. Drawing on project documentation, practitioner observation, retrospective interviews and secondary material, the paper develops and applies a four-dimensional analytical framework integrating first- and second-order cybernetics with trust research. Trust emerged as a dynamic governance achievement shaped by four interrelated dimensions: the balance between trust and control, trust asymmetry in heterogeneous consortia, feedback architecture and iterative trust calibration, and the rebuilding of trust at technical and organisational step-changes. Trust did not substitute for control but enabled viable coordination when formal governance could not absorb project complexity. The single-case design in a European defence aviation context limits generalisability, but the cybernetic framework offers transferable concepts for examining how trust, control, feedback and vulnerability co-evolve in other collaborative complex system design settings, including artificial intelligence (AI) enabled, autonomous and safety-critical socio-technical systems. Project managers and funders should treat governance design, feedback architecture, demonstrators and review points as active trust interventions that make evolving vulnerabilities salient, recalibrate the balance between trust and control, and support renewed trust-building at major technical and organisational step-changes. The paper contributes a cybernetic explanation of trust in collaborative complex system design by linking first-order concerns with feedback, control and requisite variety to second-order concerns with observation, interpretation and meaning-making. It develops an analytical framework for examining trust dynamics and shows how inter-partner trust and trust in the emerging system co-evolve over time, revealing vulnerability as a condition of complexity that can be distributed and managed but never eliminated.